For roughly forty years, "audio guide" meant a piece of hardware on a lanyard — first a cassette, then a CD player, then a smartphone-shaped Android device locked to one app. Until about 2020 that model made structural sense: not every visitor carried a capable phone, the network coverage inside a stone building couldn't be assumed, and the museum had to control the device to control the experience. Each of those constraints has now relaxed past the point of mattering for most institutions.
This piece is the honest comparison: what BYOD actually delivers, what a handset fleet actually costs, where the comparison flips, and what to do about the visitors a phone-based model genuinely doesn't cover. I'm the founder of Convo, a web-based platform in this category, so I have a position. I've tried to write the piece I'd want before signing a hardware contract or canceling one.
At a glance
| Dimension | Rented handsets | BYOD (visitor's phone) | |---|---|---| | Hardware capex | £80–£300 per device × fleet size | $0 | | Annual upkeep | Low five figures per year for a mid-size fleet | $0 | | Sanitation labor | Roughly a quarter of an FTE at two minutes per use on a busy fleet | $0 | | Visitor onboarding | Queue at the desk, ID exchange, deposit | Scan a QR code on the wall card | | Accessibility stack | Whatever the vendor shipped | OS-level (VoiceOver, TalkBack, dynamic type) | | Languages available | Whatever was loaded onto the device | Whatever the web app serves | | Coverage gap | Visitors who refuse the handset | Visitors without phone or battery | | Visitor preference | Declining since 2020 | Default among under-50s |
The rest of this piece unpacks each row and ends with the cases where rented hardware still wins.
Do museum visitors actually have smartphones?
Yes — at rates well above what's needed for BYOD to be the operating default. The Pew Research Center's most recent national survey (5,022 US adults, fielded February–June 2025) puts overall smartphone ownership at 91%. The breakdown by the demographics that show up in galleries is the part that matters: 97% of adults 18–29, 96% of adults 30–49, and 90% of adults 50–64 own smartphones (Pew Research Center, 2025).
Income raises the floor further. Among adults in households earning $70,000–$99,999, ownership is 96%; in households over $100,000, it's 97%. College graduates sit at 96%. The visitor profile most museums optimize for — urban, mid-career or older, college-educated, household income above the median — is functionally at saturation.
The honest gap is at the edges. Among adults 65 and older, ownership is 78%; for low-income seniors specifically, it falls further. International visitors arriving without a working data plan are a separate category. We'll come back to both at the end of this piece — they're real, and they're the reason a small loaner fleet is the right answer, not the absence of BYOD.
What does a handset fleet actually cost to run?
A handset fleet is not a one-time capital purchase. It's a recurring operating line, and the operating line is the part most museums underestimate. For a mid-size museum with roughly 16,000 audio guide uses a year, the annual operating cost of a fleet typically lands in the low five figures, composed of:
- Cleaning and turnaround: about 2 minutes of staff time per use, which on a busy fleet absorbs the equivalent of roughly a quarter of an FTE on cleaning alone.
- Batteries and consumables: a few hundred to low thousands per year depending on fleet size and device generation.
- Breakage and loss: plan for low-double-digit-percent of the fleet per year on most high-traffic deployments.
- Charger and rack amortization: a small but real line over the 2–3 year refresh cycle.
That's before the upfront capex on the devices themselves, which on the low end runs into the low hundreds of dollars per unit and on the high end into several hundred per unit depending on form factor. A fleet of 200 devices represents tens of thousands of dollars of capital sitting on a charging rack, depreciating against the next vendor refresh. We dig into the full five-year picture in our museum audio guide total cost of ownership piece.
The hidden line is sanitation labor. Two minutes per use sounds trivial; multiplied across the year it's a third of a full-time equivalent. Since 2020, no visitor-services director I've talked to has been willing to defend cutting that step. The labor is real and it isn't going back down.
A BYOD program replaces all of this with a printed wall card and a web URL. The marginal cost of an additional visitor is zero. The marginal cost of an additional language is also zero — the device the visitor brought already supports every script you publish.
Do visitors actually prefer using their own phones?
Yes — strongly, and the preference is asymmetric across generations. Post-2020, three behavioral shifts have stacked on top of each other.
First, hygiene. Visitors became measurably more reluctant to handle shared electronics, and the reluctance didn't reset. Several museums I've spoken with reported handset checkout rates falling 30–50% from 2019 baselines and never recovering, even as overall attendance returned. The handset is a friction asset now in a way it wasn't six years ago.
Second, the queue. The 90-second exchange at the audio guide desk — ID, deposit, lanyard, brief instructions, the visitor learns the menu — is invisible when it works and operationally expensive when it doesn't. Cutting that interaction does more for opening-hour throughput than most floorplan changes.
Third, generational defaults. Visitors under 40 expect to scan a QR code. They don't expect to learn an unfamiliar device's menu system, hold a piece of plastic to their ear in a public space, or return hardware to a desk on the way out. They will quietly skip the audio guide entirely rather than pick up a handset they don't want to use. The collapsed native-app adoption rates that cross-industry research keeps reporting in the low single-digit percentages (Frankly Green + Webb) are the same preference signal expressed differently: visitors will not install software for one visit, and they will not pick up hardware for one visit. They will scan a QR code, because scanning a QR code is a behavior they perform every week.
Does BYOD work for international visitors and seniors?
Mostly yes, with a handful of failure modes worth planning for explicitly. The four real edge cases:
- Senior visitors without smartphones. About 22% of adults 65+ don't own one. In practice, this group also tends to attend with a companion who does, or to prefer printed interpretation, or to take a docent-led tour — but the fraction who want audio and cannot use BYOD is genuine.
- International visitors without a working data plan. Roaming has gotten cheaper and eSIM coverage broader, but a non-trivial share of inbound tourists arrive with no cellular data and aren't going to enable an expensive plan to listen to a tour.
- Dead-battery and forgot-the-phone visitors. A real and recurring tail. Most galleries see at least a handful per day.
- Visitors who prefer to keep their phone in their pocket. A philosophical minority — sometimes for hygiene, sometimes for screen-time reasons, sometimes because they want a different mode of presence in a gallery. Small but persistent.
The right response to all four is the same: keep a small loaner fleet at the front desk — typically 10–30 devices for a mid-size institution, not 200 — preloaded with the same web app any other visitor would scan into. Treat them as an accommodation, the way you treat large-print labels or accessible seating, not as the primary delivery channel. The fleet cost falls by an order of magnitude. The cleaning labor falls proportionally. The visitors who need a device get one; everyone else uses what they already brought.
There's also a connectivity prerequisite worth being honest about. BYOD only works if visitors can reach your web app inside the building. For most modern museums this means publishing tours over public Wi-Fi, or designing the web app to work offline after a first load, or both. A gallery wing in the basement of a stone building with no cellular signal and no Wi-Fi is the one place rented hardware retains a clean operational advantage — see the next section.
Where do rented handsets still win?
Four specific cases. They're real, and a serious vendor will tell you about them.
1. Named-voice, high-production tours where the hardware is part of the curatorial offer. If you've cast a named voice actor, recorded a multi-stop dramatic narrative with sound design, and the audio guide is itself an institutional artifact — think the Met's audio program, or a Hirshhorn collaboration with a sound artist — the handset's lanyard and the front-desk interaction can be part of the experience. The tour is a thing the visitor checks out, not a thing they scan. Phone-based delivery doesn't ruin it, but it doesn't add anything either. The comparable traditional vs AI audio guide question has the same shape: when the production is the product, the legacy model holds.
2. Dedicated accessibility hardware. Some visitors benefit from a device with a tactile keypad, larger buttons than a touchscreen, a wired wand for hearing-loop coupling, or a screen designed specifically for sign-language video interpretation. Modern smartphones do most of this in software for most visitors, but not for everyone. If your accessibility program depends on a specific piece of hardware your visitors don't carry, you need to provide it — and that's a fleet, however small.
3. Galleries with no cellular signal and no Wi-Fi. A growing share of phone-based platforms support offline playback after a first load on the web app, but the prerequisite is getting the visitor onto a network at the entrance. If you genuinely cannot get a signal to the front of the building, or your gallery footprint is too large for a reasonable Wi-Fi build-out, rented handsets preloaded with content remain the cleanest answer. This is a smaller set of museums than it used to be, but it's not zero.
4. Mid-contract on an existing hardware-and-content deal. If you're three years into a five-year contract with a legacy hardware-and-content vendor, the migration math usually doesn't pencil until renewal. Ride it out, capture analytics, and plan the BYOD transition for the contract boundary. The right time to do this decision is not eighteen months before the contract ends.
If none of those four describe you, BYOD is the right operating default, and the harder question is which platform.
What about platform choice — native app or web app?
Don't ship a native app for visitors. Use the web. This is the most consequential implementation detail of a BYOD program, and the one most often gotten wrong.
The asymmetry between web and native app adoption is enormous. Years of cross-industry research on cultural-organisation apps — most notably the long-running tracking by Frankly Green + Webb — converge on the same finding: the average museum app pulls visitor adoption in the low single-digit percentages, and actual usage (after uninstalls and unopened installs) is lower still. A native app forces every visitor through the App Store or Play Store gate, asks for permissions, eats storage, and provides exactly one visit of value.
A web-based BYOD app — a Progressive Web App reached via QR code — bypasses all of that. The visitor taps the camera, points it at the wall card, taps the notification, and is in the app inside ten seconds with no install step. Every serious platform in this category supports this model. If a vendor is steering you toward a native-app-only delivery, that's a red flag worth pressing on.
How does BYOD change the rest of the operating model?
It collapses three line items and makes a fourth one possible. The line items that fall away: hardware capex, sanitation labor, and front-desk handset operations. The line item that becomes possible is the one most institutions previously couldn't afford — broad multilingual coverage.
A handset has to be loaded with content before the visitor picks it up. Loading a tenth language onto the fleet means re-loading every device, plus the production cost of the tenth language itself. The marginal cost of "speak to every visitor in their first language" was always high under the rented-handset model. Under a BYOD model the marginal cost of language coverage is determined by the platform, not the fleet — and on AI-narrated platforms in particular, the marginal cost of adding a language to a tour is near zero. The structural change in audio guide economics that BYOD enables is, in practice, the multilingual change. We cover that math in detail in the AI audio guide vs traditional comparison.
The other practical change: data. A handset fleet returns essentially no analytics — checkout count and breakage rates, and that's it. A web-based BYOD platform returns the same kind of behavioral data any web product does: stop completion rates, dwell times, language selections, drop-off points. Curators get a feedback loop they didn't previously have on which interpretation actually lands.
What's the right next step?
If you're currently running a handset fleet, the move isn't to scrap it tomorrow. It's to run a parallel BYOD pilot for one tour or one wing — same content, same languages, both delivery modes side by side — and compare actual checkout, completion, and visitor feedback against the existing baseline. Most platforms in this category, including Convo, offer a no-time-limit free pilot tier for exactly this purpose. Six weeks of comparison data is more useful than six months of vendor calls.
For the full category context, see the visitor experience pillar guide. For the total cost of ownership math on either model, see our museum audio guide TCO breakdown. For the broader question of how AI-narrated tours change the math on multilingual reach, see AI audio guide vs traditional.
FAQ
The verdict
For most museums in 2026, the comparison is no longer between two roughly equivalent delivery models. It's between the model your visitors already prefer and a fleet your operations team would rather not run. A handset program that made structural sense in 2010 has become an expensive accommodation for a shrinking minority, and a building-wide rental program is the wrong shape to handle that minority. The right shape is BYOD as the default, a small loaner fleet at the desk, and an honest exception for named-voice productions, dedicated accessibility hardware, no-signal galleries, and mid-contract obligations.
If you want the full visitor-experience picture before changing your delivery model, the pillar guide is the place to start. If you're ready to look at what BYOD costs in practice for your institution, our pricing is published in full and the pilot tier is free.
About the author
Eric Duffy is the founder of Convo, a platform that lets museums and cultural institutions publish multilingual audio tours their visitors can have a conversation with. He writes about visitor-experience operations from inside the category — drawing on RFP data, conversations with visitor-services directors, and the operating economics of both rented-handset programs and BYOD web-based delivery. Reach him at eric@convo.app or on LinkedIn.